Log Buffer #10: A Carnival of the Vanities for DBAs

Welcome to the tenth edition of Log Buffer, the weekly compendium of blogs for DBAs. Let’s dive right in.

Some like it gooey… or GUI anyway. Mike Hillyer lets us know about a new single GUI tools manual from MySQL. It’s a very thorough 148-pages long. I notice that it is not published under an open license (e.g. the GFDL or a Creative Commons one), and I wonder why. Does anyone out there know the answer?

In other GUI news, Matt announces that version 2.3.1 of the Kettle database/BI application for MySQL has been released. There will be a “webinar” about Kettle on September 19th on mysql.com, he adds.

Last week, several blogs covered the open-sourcing of yet another MySQL you-know-what, the popular SQLyog. On Ramblings, Stewart Smith mentioned that he is successfully using Wine to run SQLyog on Ubuntu Linux, and Webyog’s Peter Laursen steps in to mention that they are considering distributing distributing SQLyog with their own ‘yogwine’ build, rather like Google does with Picasa and Google Earth. Most obliging!

I always like blog items on the finer points of the database world, and there are a couple good ones this week.

Chris Foot continues his wonderful “The Non-Technical Art of Being a Successful DBA”, offering some miscellaneous tips and hints, like the gnomic Foot Rule of Thumb: you need to experiment and create your own rules of thumb. It’s a great series — even better than Lost!

On OLAP BI IM Stuff, Duncan Lamb points to an article in DMReview called
“Seven Habits of Highly Effective DBAs”.

Robert Treat on the zillablog offers his slides from his presentation on pl/php at the php/db|works conference in Toronto, which closes today. pl/php is a PHP-based alternative to the PL/PgSQL procedural language for PostgreSQL.

If you use Oracle and prefer angle-brackets to dollar-signs, Lewis R Cunningham introduces PL/XML – XML Based Scripting for PL/SQL, on An Expert’s Guide to Oracle Technology.

Let’s talk about Ingres, Postgres’s older sister and darkhorse of the commercial DBMS race. Dave Dargo of Ingres.com has a couple of items about rPath’s and their Icebreaker product, a database software appliance that runs Ingres on Linux. One offers his thoughts on what he calls, “the server vs. service conflict” (do you really care about machines and their operating systems?); the other, some observations on the response to the project.

On IBM’s Informix Application Development, Guy Bowerman illustrates another approach to the server/service question with a brief item about the advantages of using CoLinux to run the LAIP stack natively on a Windows machine.

Steve Muench of Dive into BC4J and ADF makes mention of the New York Metro Area Oracle User’s Group Meeting. This will take place on September 21st, and it happens to be a very full day, including five tracks, a breakfast, and a cocktail reception. They really do it big in NYC. If you’re anywhere nearby, it would be worth attending.

William Robertson looks back on a year of WTFs (Wrongs, Trip-ups, and Flubs) and other highlights on Oracle WTF, one of which, I’m flattered to note, was my invitation to do a Log Buffer. I think any of their contributors would do a good job with LB, but I fear he might demur on the grounds that LB is for DBAs and WTF is for developers, and never the twain shall meet. I replied to this, as did others, and the discussion continues. There is a difference of opinion on where the duties of one and the other begin and end. As a relative newcomer to the world of databases, I’m surprised at this DBAs-vs.-developers thing, as it seems natural to me that people practicing such closely-related disciplines within one field would have a lot to share. Go read the article, and have your say.

Jay Pipes needs some feedback on a couple things this week. One is a call for comments and input on MySQL data-auditing functionality to appear in a future version.

The second is a call for papers for the MySQL 2007 Conference and Expo. He also provides an explanation of the conference’s change of name.

vaporware inc. has a piece by Domas Mituzas on TCP tuning for your database, on Linux specifically. The point is that these systems comprise a lot of layers, and we may find that tuning outside of the DBMS itself may be required.

This dovetails with a piece by Pythian’s Paul Moen about shmmax and shmall, a couple of Linux kernel memory settings that can impact the performance of your database systems.

Karl about the Oracle Database’s Carl Reitschuster tips off his readers to a quirk of Oracle’s DBA_OBJECTS View: it does not support all object types.

On oracleblog, Robert Vollman has some observations about the NUMBER type in Oracle. Robert uses it strictly and often chooses VARCHAR2 where others might use NUMBER. He wants to hear about others’ practices.

Jon Stephens has also looked into the darker corners, of MySQL clustering in this case. On his My So-Called Blog, he briefly describes what he found there, and points to the Cluster Utility Programs section of the MySQL Manual which he authored.

Giuseppe Maxia, the Data Charmer, directs readers to a reprint of his MySQL Sandbox tutorial, which first appeared on O’Reilly Databases. Giuseppe is also going to edit Log Buffer #12 on September 29. Thank you, Giuseppe!

A late addition, courtesy of Slashdot: PostgreSQL Slammed by PHP Creator Rasmus Lerdorf in an interview with internetnews.com. Not surprisingly, that headline is an exaggeration. It seems more like Lerdorf explicitly endorses MySQL over Postgres. Then again, maybe that is a slam?

Thanks for reading. This being the landmark tenth Log Buffer, I’d also like to thank all of our contributors so far:

• Bill Thater
• Doug Burns
• Mike Kruckenberg
• Craig Mullins
• Daniel Schneller

Thanks guys, you help make Log Buffer a fun and wide-ranging look at database blogs.

If you’d like to join this illustrious cadre of editors, read the Log Buffer homepage and get in touch.

See you next week!